Innovation: Clever multi-imaging using a one-lens, one-camera solution

Go to AD-080GE

JAI's AD 080GE steps into the breach

Above: Jai AD-080GE.
Lens not supplied with camera

Generally most situations where imaging is required need one camera operating in the visible spectrum (roughly 400 - 700 nm). Sometimes however an application needs to utilise both the visible and near infrared spectrums at the same time and on the same subject. Often in these situations the application requires that for each pixel both the visible and near-infrared property is measured from exactly the same location - pixel correlated. Traditionally this is done using two separate cameras. This proves to be very cumbersome, difficult to integrate and expensive in time and budget dollars utilised. Two cameras with separate lenses and different optical axes make it virtually impossible to get perfect pixel matching between visible and infrared images for all pixels and so is a far from ideal solution.

Jai's AD 080GE one-lens, two-sensor camera has stepped into the breach and combined two cameras into one, while at the same time overcoming many of the disadvantages of using two cameras. The AD-080GE has made pixel correlated visible and Near IR capture affordable, reliable and easy-to-implement by producing concurrent images in both visible and near infrared spectra in pixel correlated alignment. More importantly, once set, this alignment remains true and does not need constant calibration.

The Jai AD 080GE's prism technology in action The image on the left illustrates Jai's special 2-CCD/prism technology for multi-spectral imaging. The prism inside the AD-080GE camera splits the light in two channels. The visible light channel captures the red, green and blue bands of conventional imaging (the normally-coloured apple at the top of the image). The near-infrared channel “sees below" the surface of organic and other materials to detect imperfections. What, superficially, looks to be merely a topical lesion on the fruit is, in fact, the commencement of decay as is seen, on the bottom left of the illustration, via the NIR sensor. (More examples below)

Previous, not-so-efficient, two-camera solution	The new JAI 2-CCD, one-camera solution
Two cameras, synchronised, imaging one object. Impossible to get pixel correlated images on non-flat surfaces and very difficult to do so on flat surfaces. .	This approach not only eliminates the alignment issues of a two-camera system, but it is easy-to-deploy and avoids the extra cost of two cameras, lenses, cables, etc.

Why is the JAI AD 080GE one-camera solution different?
Two progressive-scan sensors are mounted to a custom-designed optical prism. The manufacturing process precision-bonds the sensors onto a customised prism optic. Jai have developed their own proprietary equipment and processes to align and mount sensors to the prism accurately and reliably and are one of the very few camera companies around the world able to do so. The visible sensor is a Bayer RGB CCD sensor and is the matching colour version of the mono sensor used on the near-infrared channel so all mechanical and spatial characteristics of the two sensors are identical. A custom designed multi-faceted prism supports the sensors and is placed in the cameras optical path directly behind the lens mount. Each channel of the prism is fitted with band-pass filters. This unique 2CCD process opens up a very wide range of capabilities and Jai have already released the AD-081GE which offers extended dynamic range with 2x mono sensors.

How can Jai's unique design benefit my situation
The AD 080GE offers pixel correlation, something 2 cameras cannot offer. It can also save set-up and operating time, equipment and cabling costs. In addition, the inclusion of GigE output means:
** increased bandwidth for higher performance
** increased versatility as there is no need for specialised frame grabbers
** longer cable lengths

Jai's AD-080GE opens up the possibility for new applications or for optimising existing applications. For example it is able to view surface properties in the colour channel while simultaneously viewing just below the surface in the near-IR channel. A typical application for this capability is the inspection of fruit and vegetables where bruising under the surface cannot be detected with visible light but is clearly viewed in the image from the NIR sensor (see above) while simultaneously discerning the colour of the fruit. The ability to correlate, exactly, pixels in both colour and NIR images is impossible to do, consistently, with two separate cameras.

Other applications include the inspection of thin-screen printed plastic packaging while concurrently inspecting the contents of the packaging; applications in textile-defect detection, aerial mapping, print quality inspection and many others. Some examples follow.

Locating the rogue bean is not easy using the visible spectrum (left). The image on the right, captured with the NIR sensor (right) clearly shows the defective item.			Blister-packaging inspection with the visible image (left) showing the surface of the packaging while the NIR sensor (right) examines the contents under the foil wrapping
Electronic circuit inspection - the visible channel (mono in this case) shows components on surface while the NIR channel shows the traces between PCB layers.			Food Inspection - the image on the left (using a visible mono channel) shows several, small lesions on the surface. However, under NIR, only one of those lesions is significant because it is sub-surface

Read more about the applications, benefits and advantages of the JAI AD 080GE or to read specifications and download datasheet